Genes - Federation of American Scientists · PDF fileGenes Participants discuss the Gene research on the genetic endowment) ect and species.,.. its Mlial~. .,,,’ ., ... By convention

GenesParticipants discuss the Gene

research on the genetic endowment) ect and

species.,..

its

Mlial~. . ,,,’ .,

An interview with L. Scott Cram, Larry L. Deaven, Carl E. Hildebrand, Robert K. Jfoyzis, and Marvin van Dilia

SCIENCE: Could you sturi b) cxplait]ing (J.]-acrlj ]thaf the \’ationalLaboratory G-me Lihral.y Project provides to scimti.ft.f stu(iing hll-man chrotnosotnes, germ, and 11!%’.4?VAN DILLA: We always have to start with an explanation becausethe name “gene library” is misleading. It conjures up a vision of astorehouse stocked with many little vials, one for each of the hundredthousand or so human genes and all neatly arranged and read> forchecking out. This vision will probably be realized in the future—thedistant future—but for the time being researchers must be contentwith something less,

What we do provide is a selection from only twenty-fourvials—twenty-four libraries. Each library contains DNA specific toone of the twenty-four types of human chromosomes. the two scxchromosomes and the twenty-two pairs of autosomes. The DNA is inthe form of a set of fragments cut by a restriction enz.yrne. So thecontents of a vial are a library only in the sense of being a collec-

tion—of DNA liagmcnts—und arc in no way labeled or organimd forretrieval.DEAVEN”: The f~agments are not truly a “gene” library ei[bcr, slncceach ma! incl udc one gene, more than one gene, a portinn of’ a gene.or no gene at ail.MOYZIS: The ~ials we mail out to users don’t contain naked DNAfragments though. There’s even more to making a chronlosomc-specihc Ii brar> than culturing cells, sorting the chromosomes fromthe cells. and letting a restriction enzyme chop up the DNA moleculesfrom the chromosomes. A library must contain enough copies of theset of restriction fragments that the user has a reasonable amount ofmaterial to work with. One could obtain that many copies bychopping up the DNA from lots and lots of sorted chromosomesfrom lots and lots of cultured cells. but that’s a hard and expensiveway to go. 1 nstc:id we put a biological amplifier to work and clone thefragments.

72

First the human DNA f~agmen[s arc eachspliced into a DNA molcculc of’ a bacterialv irus—a bacteriophage. or phagc fbr shor[,Then these recombinant DNA molcculm arceach wrapped up in the bodily tl-appings oflhe phagc, and Ihc phages arc allowed toi n f e c t s o m e unfor[unatc bac[cria. Thebacteria then proceed to work [hernsclves [odeath making marry duplicate ~iruscs, eachcontaining a human DN.A fiagmcn[ in a ~ira[DNA molcculc. We gnthcr up the plaques. orcolonies of \iruses, suspenci Ihcrn i n someliquid, and ship them ofl’, So a Iibrar> actu-all~ consists of’ buman [}N’A fragments en-cased in bacteriophages.HILDEBRAND: The pbagc bocfics makeconvenient, pro[ectiw packages for the frag-ments. ,And more important. if a user wantsto work onc at a ~imc with members of’(hc setof restriction fragments i n a 1 ibral-~ —andthat’s usualll the case—all bc ncds to do ISinfect some more bacteria with the phagcs Inhis library and pick of onc plaque at a time.SCIEYCE: ffott [Y s[d u /i/vu)T I(\(Yl:)HILDEBR.AND: It depends on what theuser bas in mind. In some cases ~hc fragm-ents ma> be invest igatccf directly--thcirbase sequences dctcrm incd, fur cxamplc-and the information picccd togclhc[- [u gi \esomt piclurc of’ t h e DNA mOICCU]C a s awhole. Y’ou see. DNA ITIO]CCUICS arc so }] ugcthat details of their organ iz.atiun and fLlnc-tioning c a n . w i t h prcscn[ technolog!. bclearned only by studying them in picccs. andpreferably not pieces gcncratcd b> landombreaks in the mokxulcs but picccs cut at Ihcsame points along each molccwlc. Tha(”swhal rcstrictiun enzymes do. and i [’s d ifhcult10 O\Jf2rSti3tC bow abSOIUIC!\ Csscnliai lhcy’ arcto molecular genetics.

CJctting back to the question. in most ap-plications a library is used as a source ofprobes, A probe is a raciioacti\<-l> or fluo-rcscently labeled DNA fragment that is usccito detect and Iocatc. in a sample of DN. Abeing invcsligatcd. portions of I)NA with acomplementary base sequence. Ot course,any fr-agmcnt in a librarj can Ix Iabclcci andused as a probe. but it ma!, not bc detectinganything ~ery exciting. The cballcngc is to

74

-...

———...—

W?zatli a Chromosome?

T he hereditary information in each cellof an organism is awwsomc in exlent,complex in organization. and beau-

tiful in the dynamics of its expression.Chromosomes are the packages housing thisinformation. Most of’ LLS think of chromo-somes as the ,k”-shaped objects viewedthrough a microscope in a high-schoolbiology course, This ftimiliar manifestationof chromosomes exists. .iowcver, only dur-ing a certain brief’ stage in the life of a cell.Since chromosomes—in particular, humanchromosomes—arc the essential raw mate-rial for the Gene Library Project. a briefreview of what is now known abou~ thesecarriers of heredity is worthw’h ile.

Located in the nucleus, each chromo-some includes. as a single molecule, somefraction of’ the CCII’S complcmen~ of IJN,Aand hence of’ the CCII’S genes. The number ofchromosomes in a normal cell depends onthe species of origin of’ the ccl], Although allorganisms of a species have the same chro-mosome number, thal number is not uniqueto the species. For cxarnplc. [hc evening bat(N~vczicciu.\ ht{nzeru/iY). the red s q u i r r e l( 7’a/)?ia.scizfrt/.\ hud.w}ticm s/wu[(NA. and thehuman all have Ihe same chromosome num-ber. The dependence of’ chromosome num-ber on species is not straightforward butexhibits a general dovmward trend wi~hprogression up the evolutionary ladder.

In higher organisms anotber ~,ariation in

chromosome number occurs: all gametes(CCIIS tbat participate in sexual reproduction,such as egg and sperm cells) have the same“haploid” number of chromosomes, and allotber cells are “diploid.” with exactly twicethat number. Human gametes, f’or example,possess twenty-three chromosomes, and allother human cells possess forty-six.

How is the chromosome number of a spc-cics maintained constant from onc cellularand organisrnal generation to another? In thecase of humans, a ncw organism begins withthe formation of an ancestral diploid cell byfusion of two haploid gametes: an egg carry-ing an X sex chromosome and a sperm carry-ing an X or a Y sex chromosome. (The neworganism is a female (male) if an X-bearing(Y-bearing) sperm participated in the fu-sion. ) other cells are tben formed by succM-sivc waves of mitosis. a type of cell divisionresulting in two diploid cells, each containingduplicates of the chromosomes of both thematernal and the paternal gametes. (The ma-ternal and paternal -versions of each chromo-some are referred to jointly as a homologouspair. ) The organism grows and develops asmitosis continues and certain of the cellsbecome differentiated in function. .At somestage or stages in the organism’s Ii fe, specialdiploid cells (germ cells) undergo meiosis. atype of cell division resulting in formation offour haploid gametes. Each chromosome inthese gametes is a duplicate of dither the

Sprirrg/S.mmer t985 LOS ALAMOS SCIENCE

Genes b}’ Mail

DNA DoubleHelix

Nucleosomes

30-nanometerChromatin

Fiber

InterphaseChromosome

Fiber

Metaphase

Chromosome

maternal or the paternal chromosome or.more usual, is some combination of the twoarrived at by exchange of hereciitary materialbetween homologous chromosomes duringmeiosis. The cycle begins again with fusionof one such gamete with another from anorganism of the opposite sex.

A chromosome includes, in addition to asingle DNA molecule. a mass of special pro-teins (histones) roughly equal to that of theDNA molecule (some 15 percent of the totalchromosomal mass), small amounts of manyother DNA-binding proteins. and RNA. Thehistones make the DNA molecule more flex-ible. by neutralizing the negative charges ofits phosphate groups, and thus play an essen-tial role in packaging the very long DNAmolecules (which, for humans, average 5centimeters in length) within the confines ofthe cell nucleus (which is between 5 X 1 ()-4

and 10“” ~ centimeter in diameter), Many de-tails of this remarkable engineering feat arcas yet unknown. but sc~cral levels of packag-ing are involved (Fig. I).

Tbe packaging begins with the formationof%ucleosomes.” eacb consisting of-a lengthof linking DNA and a length of DN4 woundaround a core of histones. N-cxt. some regularpacking of the nuclcosomcs, also in~olvinghistoncs, Icads to the structure known as (be30-nanometer cbromatin fiber (tbc doub]chelical strand of DNA itself is about 2 nano-meters in diameter). This fiber in turn isorganized into a sequence of “looped do-mains,” each of- which is thought to be ac-tivated as a unit during gene cxprcssioo. Thesequence of looped domains is ftlrthcr con-densed in some fasbion to fcmm the “inter-phase” chromosome, portions of which re-condense and recondense as various genesarc activated and dcactitatcd during bio-synthesis by the cell.

The ultimate in DNA packaging occurs asa germ cell prepares for meiosis or as asomatic cell (any ccl] other tban a germ cell)prepares for mitosis. Then each interphasechromosome is duplicated, and the two Fig. 1. Many details of the packaging of a DNA molecule and various proteins intoidentical “sister chromatics. ” which arc the entity known as a chromosome are yet to be determined. These schematicjoined at some point along their lengths (the illustrations are merely suggestive of the mechanisms and configurations that may“centromere”). form a longitudinally syn- be involved.

1 1

Centromere\ I

Sister Chromatics

LOS ALAMOS SCIENCE Spring/Summer 1985 75

Fig. 2. Pltotomicrograph of metaphasechromosomes from human leukocytes(white blood ceils). The cells were culturedfor 48 hours, blocked at metaphase by ex-posure to the drug Colcemid for 2 hours,swollen in a hypotonic solution, and treatedwith a fixative. The chromosomes were re-leased from the metapitase cells simply bydropping the swollen cells onto microscopeslides, which breaks the cellular mem-branes. (The dark, round objects are nucleireleased from interphase cells.) Thechromosomes were then partially digestedwith ttypsin and treated with Giemsa stain.Since trypsin #referentially digestschromosomal proteins bound to DNA richin guanine-cytosine pairs, those portions ofa chromosome containing such DNA areless readily stained. The result is a patternof heavily stained (dark) and lightlystained bands along the length of eachchromosome. The patterns of these bandspermit unambiguous identification ofchromosomes.

t I

3 4 5I

Fig. 3. A display of the metaphasechromosomes of an organismj banded as inFig. 2 and arranged in order of decreasingsize, is known as the karyo~pe of the or-ganism. Shown here is the karyotype of ahuman male. By convention the pair of sexchromosomes is displayed separately fromt h e homologotis p a i r s of autosomes.Chromosomes are designated by a size-related number and, historically, by amorphology-related group: group A in-cludes numbers 1 through 3; group B,numbers 4 and 5; group C, numbers 6through 12; group D, numbers 13 through15; group E, numbers 16 through 18; groupF, numbers 19and20;group G, numbers 21and 22. The similar morphology of themembers of a group made them extremelydi@cult to distinguish before discovery in1970 of the banding method described inFig. 2.

I

6 7 8 9 10 12I

,

15

,)

t I

-10 ~m

76 Spring/Summer 1985 LOS ALAMOS SCIENCE

Genes by Mail

WHAT I S A C H R O M O S O M E ’ !

metric double structure with the familiar Xshape. These “metaphrase,” or ‘“mitotic, ”chromosomes are easily visible through anoptical microscope (Fig. 2) and were firstobserved by German CCI1 biologists in theearly 1880s.

The metaphase chromosomes of a givenspecies can be distinguished on the basis of anumber of properties, such as overall length,DNA content, or location of the centmmere.(Except in the case of an X-Y pair, thedifferences between the two members of ahomologous pair of chromosomes are muchless pronounced than the differences amonghomologous pairs.) The most definitiveidentifying property is the pattern of bandsproduced on metaphase chromosomes byselective digestion of the chromosomal pro-teins and subsequent staining. Figure 3 dis-plays such banding patterns for the humanchromosomes.

Which genes are packaged in whichchromosomes, and exactly where, is knownonly for some 700 of the estimated 100,000human genes. Examples of those for whichthis information is known include the genefor Huntingtons’s chorea (chromosome 4),for interferon (chromosome 9), for $globin(chromosome 11), and for hemophilia andred-green colorblindness (the X chromo-some).

The marvelous choreographies of mitosisand meiosis normally transmit intact andunchanged the time-tested chromosomes ofan organism from one cellular and organ-ismal generation to another. But sometimeserrors. such as rearrangements of chromo-somes or departures from the normal num-ber, do occur. The consequences of errors inmeiosis are usually grave, if not fatal, to theorganism that inherits them. For example, anabnormal number of chromosomes arefound in the cells of about half of the humanembryos that are spontaneously aborted, andthe presence of three copies of human chro-mosome 21 results in Down’s syndrome.Errors in mitosis also have serious conse-quences, as evidenced by the association of

chromosome rearrangements with varioustypes of cancer. ■

L05 ALAMOS SCIENCE Spring/Summer 1985

find the interesting probes, ones that revealsomething unusual about the DNA, such asan alteration related to a genetic disease or amutation.SCIENCE: Have libraries been producedbefore?CRAM: Whole-genome libraries, ones con-taining DNA fragments from all the humanchromosomes, are already available, and soare a few chromosome-specific libraries, onesfor the chromosomes that are easier to sort.But this project will bc the very first toprovide chromosome-specific libraries for alltwenty-four types of human chromosomes.VAN’ DILLA: With a chromosome-specificlibrary a user can begin his research byfishing around among a pool of DNA frag-ments from a particular chromosome ratherthan the whole sea of human DNA. This is atremendous advantage, and the librariesshould increase the productivity of researchin molecular genetics just as separatedisotopes increase the productivity of researchin nuclear physics,DEAVEN: I have some data to support thatexpectation, at least as far as one applicationof the libraries—gene mapping—is con-cerned. By gene mapping is meant localizinga gene on a particular region of a particularchromosome. In 1958. when molecular gene-tics was still in its infancy. only about fourhundred human genes had been identified.These had been classified as dominant orrecessive and autosomal or X-linked but.except for the X-linked genes, could not bemapped. That had to await, first of all, thediscovery some ten years later of chromo-some banding, a technique for unambigu-ously distinguishing all twenty-four types ofhuman chromosomes. This discovery madeit possible to map a limited number of genesby associating chromosome abnormalitiesand genetic diseases. AI about the same timehuman-rodent hybrid cell lines becameavailable, and information about the pro-teins synthesized by various of these celllines, each carrying only a few humanchromosomes, made it possible to map moregenes. Still, by 1973 only 64 genes had beenassigned to particular autosomcs and 15510

the X chromosome. By 198 I the total num-ber of mapped genes had increased to onlyabout four hundred and fifty, but in that yearmolecular genetics provided a powerful newtool for mapping. It was discovered that whatare known as restriction fragment lengthpolymorphisms could pinpoint the locationsof genes about which practically nothing isknown except the observable evidence oftheir expression, such as the symptoms ofdiseases.

A restriction enzyme cuts a DNAmolecule into fragments of various lengthsby breaking bonds at every occurrence of acleavage site, a certain sequence of bases.Mutations can generate new cleavage sitesand destroy existing ones, and these changescause differences in the sets of restrictionfragments from different individuals. For ex-ample, in the fragments produced by therestriction enzyme }fpaI from the DNA ofAmerican Blacks, the (3-globin gene usuallyappears on a 7.6-kilobase fragment but some-times on a 7.0- or 13.O-kilobase fragment.Differences like these arc called restrictionfragment length polymorphisms, or RFLPs..An RFLP can be detected by using clcc-trophoresis to separate, according to length,the restriction fragments from, say, two in-dividuals and seeing whether a probe thatincludes the RFLP sticks to different placesin the two lineups.

An RFLP within a gene is obviously in-herited along with the gene, but so also is anRFLP so close to a gene that it is not sepa-rated from it during meiotic recombination.So the location of such an RFLP on a chro-mosome is a very close approximation tothat of the gene. This method of mapping,say- for a gene that causes some disease,involves testing many probes to find one thatreveals an RFLP unique to people with andwithout the disease—that’s the hardpart-and then mapping the probe on thechromosome—that’s the easy part.

Since 1983 about three hundred RFLPprobes for genes have been identified andmapped per year. Even that rate, though, ispainfully slow considering how many moregenes remain to be mapped. But our chromo-

77

some-specific libraries will make the searchfor probes less difficult, and we estimate thatabout two thousand RFLP probes for geneswill be identified and mapped in the first yearthe libraries are available, and even more insucceeding years.HILDEBRAND: Once a probe for, say, adefective gene has been identified, it hasapplications in addition to gene mapping. Itcan be used to detect the gene in unbornchildren and in unknowing carriers of thegene—unknowing because the gene may berecessive, not clearly expressed. or, like thegene for Huntington’s chorea, expressed onlylate in life. People from families with his-tories of genetic disease usually want suchinformation to help them make decisionsabout the course of pregnancies or of theirown lives,

RFLP probes for genes are also veryvaluable to basic research in molecular ge-netics. For instance, by studying the in-heritance patterns of probes for variousgenes, we can learn about the relative loca-tions of the genes along the DNA molecules,SCIENCE: What is ~he history qftkeproject?

DEAVEN: About three years ago a group ofus in the Life Sciences Division at Los Ala-mos decided to look into the possibility ofestablishing a new program oriented aroundthe structure and functioning of mammalianchromosomes. We envisioned a small groupof related projects to be supported by somefederal agency, Among the ideas we came upwith, that of producing DNA libraries foreach of the human chromosome typesseemed particularly worthwhile and, fur-thermore, related to two establishedstrengths of the Laboratory -GenBank,which is a repository for DN,A sequence data,and the National Flow Cytometry Resource,which organizes our capability in the tech-nology required for sorting chromosomes.

The DOE’s reaction to our idea was quitefavorable. They thought that the librarieswould provide the means for new researchcapabilities at the national laboratories andelsewhere and that production of thelibraries would be a unique contribution tothe biomedical community. Of course, onc

L. Scott Cram of Los Alamos, abiophysicist, is a specialist in the ap-plication of flow cytometry to identi-fication and sorting of chromosomes.His research interests include the roleof karyotype instability in tumori-genesis.

ultimate payoff of the project will probablybe a better method of detecting mutations inhumans caused by exposure to radiation orother mutagens, and this is a basic interest ofthe DOE’s Office of Health and Environmen-tal Research. In any case, the DOE decidedto fund the project as a service to the biologi-cal and medical communities. [n discussionswith personnel at the DOE, the suggestionwas made that we initiate the project jointlywith a group of biomedical scientists at Liv-ermore, which also was strong in flowcytometry,

People at Livermore were enthusiastic,and together we began by meeting, in Octo-ber and November of 1983, with a group ofmolecular biologists and human geneticistswho advised us about what was needed andthe relative priorities to assign to thoseneeds. This advisory group suggested tbat wc

construct, as quickly as possible, a set oflibraries to be used as sources of potentialprobes for gene mapping and diagnosis ofgenetic diseases. For those applications smallDNA fragments are best, since a small frag-ment that includes a sequence coding for agene is not likely to include in addition oneof the many noncoding repetitive sequencespresent in DNA, So that’s what we’re doingnow—making libraries of relatively smallfragments. When these Phase I libraries arecompleted, we’ll go on to constructing thePhase 11 libraries, with larger human DNAinserts, to be used for basic research in genestructure, arrangement, and expression.

This is a good opportunity for me toacknowledge the foresight of our technicalrepresentatives at the DOE. Usually, tbemost important products of research are

Larry L. Deaven, a c}~togeneticist, isprincipal investigator for the GeneLibrary Project at Los Alamos. Hisresearch interests include the chromo-some constitution of evo[ving cellpopulations and the chromosome dam-age induced by agents associated withenergy production.

78 Spring/Summer 1985 LOS MAMOS SC1t?NCt?

Genes b}) Mail

Carl E. (Ed) Hildebrand of Los Ala-mos, a biophysicist turned molecularbiologist, is a specialist in recombinantDNA technology. His research inter-ests include DNA replication, chro-matin structure, and gene expres-sion.

publications, and to some extent research isfunded on the basis of the cxpectcd numberand quality of’ the publications. But for thisproject the DOE set tbat criterion aside infavor of a different goal—the production oflibraries, Although we expect importantpublications f’rom our work on the project,our major aim is 10 produce high-qualitylibraries for use by interested scientiststhroughout the world.

In addition, the Gene Librar> Projectwouldn’t be as f’ar along as it is if’ it badn’tbeen funded b> the DOE. which can ac~ \cr},quickl} when somc[hing regarded as worthsupporting comes along, We have now com-pleted approximatcl! 75 pcrccnt of-the Phase1 libraries. If our technical representatives atthe DOE had not ufiders[ood the significanceand timeliness of the project. wc might well

still be in the organizational stages.

LOS ALAMOS SCIENCE Spring,/Summer 1985

SCIENCE: How does it happen {haf twoweapons laboratories have the personnel andfacilities required, for such a project?DEAVEN: That’s not so much of ananomaly as you might think. Tbe AEC had anatural interest in the biological effects ofradiation and realized. as have its successors,that understanding those effects in detail hadto be based on knowledge of cellular proc-esses at a molecular le~el. So in addition toprogrammatic researcb on the biologicaleffects of radiation—and now of’ variouschemicals associated with other means ofproducing energy—basic research in cellu-lar biology has received strong support at thenational laboratories since their begin-nings.

SCIEh-CE: J? “ha~ L’[nd (/f (()//u/I()t(IIi(ItI /z-tueen [he two [al)oratorim i.j i)!iolw,d, utldho)t well is (t w(wkitl,y’

Robert K. Moyzis, a molecularbiologist, is Leader of the GeneticsGroup at Los Alamos. His researchinterests include the organization ofhuman DNA and the regulation ofgeneexpression.

Marvin Van Dilla, aprincipal investigator

biophysicist, isfor the Gene

‘Libra~ Project at Live;more. His re-search interests include the develop-ment of flow instrumentation and itsapplication to biological problems.

\’AN DILLA: W’c collaborate in the sense of’drawing on each other’s expertise rather thanin lhc sense of parceling out \arious aspectsof the project to one or the other laboratory.Both of us have gone [hrougb the wholeprocess of’ making Pbase I libraries. but ourproducts are not icicntical—by edict ratherthan b} chance. The Livcrmorc Ii brarics con-tain fragments of DNA cut by onc restrictionenz>, mc. and the Los Alamos I ibraries con-tain fragments cut by a difcrcnt restrictionenzyme. So a fragment missing from onctcrslon, because the recombinant DNArnolccule containing that fragmen[ is toolarge to fit into the bactcriophagc, will quitelikely bc i“uund in the other.

In my judgment the collaboration hasbeen ver! cffectivc, with each Iaborator} con-tributing particular strengths. For cxarnplc.we at Ll\crnlorc had greater cxpcricncc in

79

sorting humancytomctm-s and

chromosomes withhad been developing

flowone

capable of doing the job at a rate compatiblewith the scope of the project, We need some-thing on the order of a million sortedchromosomes for each library. (-em -merciall~ alailablc flow cytometcrs can sortthat many in a few tens of hours-anci that’soperating time. so the job actual l>’ takesmuch Iongcr—but the high-speed sorter re-duces the sorting time to just a fcw hours. WCmade copies of the basic components of oursorter and sent them here. 1 understand theLos Alamos machine is now up and running.

.AS for Los .Alamos. the> brought to theproject greater expcricncc and expertise inrecombinan~ Dh” A technology and in facthad made a fiw chromosome-speci f iclibraries for lhc Chinese hamster before theproject even started. The Los 41amos grouphad also explored the use of h>brid CCIIS assources of human chromosomes.

Our first joint meeting, at Los .Alamos inDecember of ’84. was proof to me of a suc-cessful collaboration. Almost e~;eryone in-volved in the project at Livermore. fromPh.D.-s to technicians. attended the meetingand took the opportunity to discuss commonproblems and exchange tricks of the tradewith their counterparts at Los Alamos. Ourchief cell farmer, for instance, had neverbeen to Los Alamos to talk to the chief cellfarmer here. They both do the same thingwith somewhat different techniques andhave developed somewhat different ways ofgetting around the many nitty-gritty prob-lems that crop up in the busines of cellculture. Together they did two experimentsduring the meeting. Onc day they grew cellsand isolated the chromosomes using thetechnique that’s favored at Livermore, andon the next day they went through the samesteps using the technique favored at LosAlamos. Each learned something from theother and seemed very happy about thewhole experience.

SC IEN’CF,: flolt” /}/1((”/1 inf(’)”c$l /?(J.\ /v’c}/

show?? it7 [Ac [ihrurics;’

80

Making the Libraries

Supplying theChromosomes

A s raw material for each chromosome-specific library, we must collect be-tween half a million and a few

million of a particular chromosome by sort-ing through tens to hundreds of millions witha flow cytometer. Our original hope was toobtain such large numbers of chromosomesfor sorting by culturing, or multiplying in

}Iifro. human cells. namely, fibroblasts fromhuman foreskin tissue. But. as discussedbelow. these cells pro~cd to bean acceptablesource for only the smaller chromosomes,and another source had to be found for thelarger.

We culture fibroblasts in three stages.beginning by placing small pieces of’ the tis-sue sample, together with nutrients essentialfor growth, in a small plastic container. (De-spite the translation of in ri[m. plastic haslargely supplanted glass as a surface for cellculture. ) The fibroblasts migrate fkom thetissue, adhere to the bottom of the container,and multiply by mitosis. (This process issimilar to that by which tibroblasts iH ri}o

repair a break in the skin. ) Mitosis stops

when the bottom of the container is coveredwith a monolayer of cells (Fig. I ), The mono-layer is then partially digested with theenzyme trypsin to yield a suspension ofsingle cells. This suspension is ~iividedamong several containers. fresh growth me-dium is added. and the culturing is repeated.again until a monolayer has formed.

Our aim is to maximize the number offibroblasts at mctaphase (on the verge ofdividing), since only those cells contain themelaphase chromosomes that can be sorted.Ob\;iously, the mom rapidly a population ofcells is multiplying, the larger is the fractionof that population at metaphase. But thefibroblasts begin to multiply less rapidly afterabout fifteen to twenty mitoses and ceasem u l t i p l y i n g ahogcther after about fifty.Therefore, after the second culture some ofthe frbroblasts are suspended in a mediumcontaining glycerol (or some other ice-crystalinhibitor) and frozen in 1 iquid nitrogen.These cells, which can be thawed and rc-cultured at any time up to several years later,scrwe as a reservoir of fibroblasts at the height

Spring/Summer 1985 LOS ALAMOS SCIENCE

Genes by Mail

.

Fig. 1. Cells in culture multiply bymitosis and spread over the surface ofthe culture dish until a confluentmonolayer of cells is formed, Shownhere is an optical micrograph of a con-fluent monolayer of fibroblasts cul-tured from apiece of human skin tissue(dark area).

Fig. 2. Cells can multiply manyfold inlarge, slowly rotating cylinders (“rollerbottles”) since a large surface area iskept in contact with the culture me-dium.


of their multiplication rate.The final culture of the remaining

fibroblasts takes place in large “roller bot-tles” (Fig. 2). When a monolayer has almostformed, the drug Colcemid is added to arrestthe cell cycle at metaphase. After exposure tothis drug for an appropriate time. the fractionof cells at metaphase increases from the usualmaximum of 0.5 percent to between 30 and50 percent. Since the metaphase fibroblasts“round up” from the culture surface becauseof their more spherical shape, they are easilyseparated by gentle shaking of the roller bot-tles.

The next step is to isolate the chromo-somes from the cells. first, the cells areswollen in an “isolation buffer” with an os-motic pressure lower than that of physiologi-cal saline. The buffer is then forced through afine-gauge hypodermic needle to break thecell membranes (the nuclear membranes arealready dissipated in metaphase cells) andrelease the chromosomes. The buffer solu-tion also contains materials. such as proteinsthat coat chromosomes or dyes that inter-calate with DNA, to stabilize the morphol-ogy of the chromosomes. Although thisisolation step sounds simple in principle.difficulties that plague even the best methodsare responsible for our yield of only 2 10 6sorted chromosomes from every 100 cells.

The final step in preparing the chromo-somes for sorting is addition to the buffer oftwo fluorescent dyes, One of these bindspreferentially to DNA sequences rich inadenine-thymine base pairs and the other tosequences rich in guanine-cytosine basepairs. The fluorescence intensities excited inthe dyes serve as the basis for flow-cytometric analysis and sorting of thechromosomes,

Analysis of chromosome preparationsfrom human fibroblasts indicated that all thechromosomes except numbers 9 through 12were sufficiently well resolved for sorting (seeFig. 2 in “Sorting the Chromosomes”). Inpractice, however, sorts for the largerchromosomes were very time-consuming,apparently because the shearing action thatbreaks the cell membranes also breaks up

82

many of the larger chromosomes and greatlyreduces their numbers. More important. thepurities of these sorts were reduced to lessthan the acceptable level of 90 percent bymany doublets of smaller chromosomes,Evidently the shearing action is not sufficientto break up these doublets. which form dur-ing isolation and cannot be distinguished bythe flow cytometer from the larger chromo-somes. Furthermore, reducing the shearingforce to avoid breaking the la rgerchromosomes only increased the number ofdoublets.

As a result, we have had to turn to human-hamster hybrid cells as a source of the largerchromosomes. Such cells result from fusionof human and hamster cells and initiallycontain a complete set of both human andhamster chromosomes. However, after sev-eral weeks to several months in culture, thehybrid cells spontaneously and randomlylose some of their complement of humanchromosomes. These losses make it possibleto clone hybrid cells carrying selected humanchromosomes. For our purpose an idealhybrid cell is one with no more than threehuman chromosomes, each separable in aflow cytometer from each other and from thehamster chromosomes.

Since hybrid cells with a small number ofhuman chromosomes are useful in otherareas of research, various cell lines have beenestablished and frozen at a number of Labora-tories. We are indeed grateful to those whohave shared their cell lines with us for theGene Library Project.

Culture of the hybrid cells is similar to,and no more difficult than, that of humanfibroblasts. However, as sources of chromo-somes for the libraries, hybrid cells are farfrom perfect, Their habit of unexpectedlylosing human chromosomes, for instance,requires frequent monitoring for the con-tinued presence of the desired chromosomes.Even more troubling is the fairly commonoccurrence of DNA exchange between thehuman and the hamster chromosomes. Suchexchanges are difficult to detect but, un-detected, contaminate the libraries with

DEAVEN: Well, we recently filled requestsfor over two hundred and fifty Phase Ilibraries from users in the United States andseven other countries. We expect many morerequests as word of their availability getsaround.MOYZIS: I would be very surprised if’ever]research group in the world working on hu-man genes wouldn’t want at least one of thelibraries, expecially after the Phase IIlibraries arc available.SCIENCE: Do you exercise any control overthe uses to which the libraries are put?DEAVEN: No. The request form for alibrary includes a question about the natureof the proposed research. but we ask thatprimarily because we want to keep a recordof the various applications. (Users funded bygovernment agencies have already agreed tothe guidelines established by the NationalInstitutes of Health for recombinant DNAresearch. and industry voluntarily abides bythose guidelines.

Incidentally the plan for the future is thatthe NIH will establish a repository for thelibraries we produce and will handle thedistribution. They will also collect informa-tion determined by the users about t h elibraries, such as purity and completenessdata and characteristics of probes isolatedfrom the libraries, This in information willserve as feedback to us for improving futurelibraries and will of course be valuable toother users. NIH also plans to establish arepository for probes pulled from o u rlibraries and others.MOYZIS: One thing we do request of usersis that they don’t pass the libraries on toother investigators. We want the libraries tooriginate directly from us, or from the re-pository when that comes about. One reasonfor this is so that wc can keep in touch withall the users. A more important reason fornot wanting the libraries passed around is topreserve their characteristics. Each ampli-fication of a library unavoidably introduceschanges. For example, since phages carryinghuman DNA fragment A do not multiply atexactly the same rate as those carrying frag-ment B. after several amplifications thc rela-

Spring/Summer 1985 LOS AL.AMOS SCIENCE

Genes by Mail

tive numbers of fragments A and B will bequite different, Some fragments may disap-pear altogether.

Another kind of change that occurs when alibrary is amplified is rearrangement of thefragments. One can start, say, with a frag-ment containing genes A, B, and C in thatorder and, because cloning is somethingmore than stamping out pennies. end up withthe genes in the order C, A, and B, which ismisrepresentative of the chromosome. It wasto minimize this kind of change that wechose the bacteriophage Charon 21A as thecloning vector for the Phase 1 libraries, sincein this phage the tendency for the humanDNA insert to get rearranged is minimal.When we move on to the Phase 11 librariesand start using other cloning vectors, we’llhave to worry more about this problem.

SCIENCE: What are some applications ofthe libraries?

VANDILLA: We’ve already discussed genemapping and mentioned diagnosis of’geneticdiseases, detection of mutations. and basicresearch in molecular genetics. Some otherinteresting applications are comparisons ofDNA sequence organization among the vari-ous human chromosomes, pedigree analysis,and comparison of human DNA with thoseof other species.

MOYZIS: Before we go on to other applica-tions, I’d like to point out that gene mappingis more than just collecting dry facts—itprovides basic clues about possible mech-anisms for controlling gene expression. Forexample, some thousands of genes are ex-pressed only in nerve cells. How can that be?If we find that these genes are scattered aboutall over the genome, we may conclude thatsome structural feature near each gene reg-ulates its expression. But if we should findthat all are located in the same region of justa few chromosomes, then a very different,global regulation mechanism may be atwork,

There’s also a wealth of data relating vari-ous pathological conditions. includingcancer, to chromosomal abnormalities suchas translocations and deletions. Why is this

LOS ALAMOS SCIENCE Spring/Summer 1985


Sorting theChromosomes

B efore the advent of flow cytometry,human metaphase chromosomescould be coarsely partitioned into at

most ten fractions by centrifugation orsedimentation techniques. Now. a state-of~the-art flow cytometer can separate thesechromosomes. with an accuracy of up 95percent and at a rate of up to 400 per second.into twenty-four fractions, one for each of thetwo sex chromosomes and the twenty-twohomologous pairs of autosomes. Originallydeveloped at Los Alamos in the late 1960s tomeasure the DNA content of cells,* the flowcytometer is absolutely essential to our goalof producing very pure and highly represen-tative chromosome-specific libraries of DNAfragments.

The necessary accuracy of sorting, andhence purity of the libraries, is achieved witha dual-beam flow cytometer (Fig. 1). Thisinstrument “recognizes” a particular humanchromosome on the basis of a property of theDNA it contains, namely, the ratio of

adenine-thymine to guanine-cytosine basepairs. This ratio, which varies from one chro-mosome to another, is translated intomeasurable fluorescence intensities by stain-ing the chromosomes with two fluorescentdyes. One (Hoechst 33258) binds with highspecificity to sequences rich in A-T basepairs, and the other (chromomycin A3) tosequences rich in G-C’ base pairs. As thestained chromosomes intercept the two laserbeams of the cytometer, the fluorescence in-tensities excited in both dyes are measuredand used to trigger sorting if they fail within acertain range of the intensities characteristicof the desired chromosome. The data col-lected during this “bivariate” analysis of achromosome preparation are monitored inreal time and displayed as a “flow histo-gram" (Fig. 2). Such a histogram is indicative

*See "Flow Cytometry: A New Tool for Quan-titative Cell Biology ’’ Los Alamos Science, Volume1, Number 1, Summer 1980.

83

(a) Hoechst 33258

to

Photo multiplier

and Electronics

for Analysis

Deflection

Plates

(b) Hoechst 33258

Fig. 2. (a) A flow histogram obtained fromanalysis of a typical chromosome prep-aration from human fibroblasts. The histo-gram is a record of the number offluorescence events (each presumably due toa single chromosome) versus thefluorescence intensities of the dyeschromomycin A3 and Hoechst 33258. Peaksin the histogram are attributed to differentchromosomes as shown. This histogram in-dicates that chromosomes 9 throngh 12cannot be sorted with the necessary ac-curacy from such a mixture of all the hu-man chromosomes. In practice, neither canchromosomes 1 through 8, X, 14, and 15(see “Supplying the Chromosomes"). Forthese chromosomes the necessary accuracyis achieved by sort ing from thechromosomes in human-hamster hybridcells rather than those in humanfibroblasts. (b) A flow histogram of a chro-mosome preparation from human-hamsterhybrid cells carrying human chromosomesX, 12, and 15. Note the excellent resolutionof the human chromosomes. Hybrid cellscarrying various human chromosomes areused as sources for those chromosomes notsortable at an acceptable purity fromfibroblast preparations. Note that thehamster chromosomes lie on a straight linein the histogram; this is evidence of a nearlyconstant ratio of adenine-thymine basepairs to guanine-cytosine base pairs in theDNA molecules of these chromosomes.


Genes by Mail

II. SORTING

of the accuracy with which the chromosomesin the preparation can be distinguished.

One problem we faced in sorting thechromosomes was finding a suitable stabi-

lizer, a material added to the isolation bufferto stabilize the morphology of’the metaphasechromosomes. Some of these materials ap-peared to decrease the specificity of the dyes(thus decreasing the sorting accuracy) and todegrade the chromosomal DNA (thusproducing fragments that cannot be in-corporated in recombinant DNA molecules),These undesirable side effects have now beenminimized by us ing the polyamidesspermine and spermidine as stabilizers.

The speed of sorting is crucial to produc-ing highly representative libraries, which re-quires cloning DNA fragments from verylarge numbers of sorted chromosomes. Forexample, about two million were needed foreach library produced during the first phaseof the project. (These Phase I libraries con-tain relatively small DNA fragments.) Withthe commercial flow cytometer then avail-able, sorting that many chromosomes tookabout twenty hours, a not unreasonable time.But ten times more sorted chromosomes willbe needed for each of the Phase 11 libraries(which will contain relatively large DNAfragments), and two hundred hours is indeedan unreasonable time. With this in mind. ourcollaborators at Livermore developed thehigh-speed sorter now in use at both labora-tories. Its approximately tenfold greater sort-ing speed is due to more rapid formation ofdroplets, a modification that permits sortingfrom more concentrated chromosomepreparations. We are also investigating thefeasibility of simultaneously sorting four,rather than two, different chromosomes bycharging the selected droplets to one of twolevels of either polarity.

The successful application of flowcytometry to sorting chromosomes isevidence of the many advances made in itspractice. But the technology still has its draw-backs. It is expensive and, more important.very labor-intensive. Computerized moni-toring and control may help reduce the hu-


so? A good first question to answer is whatgenes are located near those abnormalities.

HILDEBRAND: The chromosome translo-cation associated with Burkitt’s lymphoma, acancer of the immune system, is known toput a cancer-causing gene near a regulatorysequence for an immunoglobulin gene. Butwe don’t know for sure whether this observa-tion has general significance, For example,are there many such rearrangements thatlead to pathological states?

DEAVEN: The same argument holds formapping close to the translocations knownto be involved in many hereditary diseases.If we had documentation of what genes werelocated on the chromosomes that participatein a translocation, then we would have in-sight into the mechanisms at work in ex-pression of the disease.

MOYZIS: It wasn’t too many years ago thatthe immense amount of DNA in a humancell was regarded as having been just throwninto the nucleus like a mess of spaghetti.Now more and more evidence is accumulat-ing that this mess has a very defined orderand that any perturbation of that order re-sults in some alteration in the individual. It’snow known, for instance, that approximatelyone-half of all pregnancies end in spon-taneous abortions at a very early stage ofembryonic development, and the chromo-somes of about one-half of all spontaneouslyaborted embryos are badly scrambled. So thegenetic diseases that we know about are onlya very small fraction of what can possibly gowrong. Most alterations in embryonic DNAlead to embryonic death.

DEAVEN: Between 2 and 3 percent of allnewborns are afflicted with one of the morethan three thousand known genetic diseases.And these numbers don’t include some rel-atively common adult disorders. such asAlzheimer’s disease, that have a genetic com-ponent. So genetic disease is a major biomed-ical and social problem. I think that thelibraries will accelerate the identification ofprobes for rapid, inexpensive prenataldiagnosis and genetic counseling. Biotechcompanies should find this a profitable

field-finding and marketing batteries ofsuch probes.

The ultimate goal, of course, is not simplyto predict the occurrence of a genetic diseasebut to uncover the biochemical defect in-volved. It’s amazing how little we under-stand about something as common asDown’s syndrome. We know that it’s as-sociated with three copies of chromosome21, but why an extra copy of that chromo-some should have such devastating effects isnot known. Down’s syndrome may be anextremely difficult disease to tackle, though,since a whole chromosome is involved. Dis-eases caused by defects in single genes shouldbe a lot easier.

HILDEBRAND: Once a disease has beentraced to a defect in a gene, copies of thenormal gene can be produced by recombi-nant DNA techniques, and this opens thepossibility of’ gene therapy, of treating thedisease by introducing the normal gene intothe patient. Gene therapy is still very muchin the experimental stage, and it’s a sensitiveissue because of the ethical and legal implica-tions of intervention with the humangenome.

MOYZIS: Yes, but probably no one is goingto complain about attempts to apply genetherapy to really terrible diseases—say,Lesch-Nyhan syndrome with its symptomsof severe mental retardation, kidney damage,cerebral palsy and self-destructive behavior.What is worrisome to many is the specter ofeugenics, of using such techniques to altermental or physical characteristics. Manypeople feel it’s one thing to genetically engi-neer plants or animals for desirable traits,and quite another to genetically engineer hu-mans. In the first place. who’s to say what aredesirable traits? I strongly believe, however,that the types of genetic tampering that arcthe most feared are exactly the ones that arcthe least likely to be performed in the im-mediate future, mainly because we don’tknow how. It’s extremely unlikely that any-one alive today will ever see genetic engineer-ing applied to so complex a trait as in-telligence. We haven’t the foggiest notion

85


Cloning the DNAThe cells have been cultured, the

chromosomes isolated and sorted.Now can begin the final act in the

making of a library—that of multiplying theDNA in the sorted chromosomes by a factorof more than a million. Some tricks of re-combinant DNA technology make such greatamplification possible, and the magicianswho perform these tricks are lowly but ac-commodating bacteria and viruses.

Briefly, four steps are involved in cloningthe DNA: preparation of DNA fragmentsfrom the sorted chromosomes and frombacterial viruses (bacteriophages, or simplyphages); formation of recombinant DNAmolecules including both chromosomal andviral DNA fragments: assembly of phagescontaining the recombinant DNA molecules:and cloning of the phages (and hence replica-tion of the chromosomal DNA fragments) ina bacterial host. For concreteness we de-scribe, and illustrate in the accompanyingfigure, the making of what we call a Phase Ilibrary, one based on relatively smallchromosomal DNA fragments. (Our Phase 11libraries, which will serve a differentpurpose, will be based on larger fragments. )

First the DNA molecules in the sortedchromosomes must be extracted from theother chromosomal components. For rea-sons that will become clear later, this extrac-tion must be performed very gently to mini-mize breaking the very long and very fragileDNA molecules.

The extracted DNA is then digested withthe restriction enzyme EcoRI, which rec-ognizes the sequence of six bases–G–A–A–T–T–C– and, as shown below,cuts the DNA at each occurrence of’ thissequence by breaking one covalent bond on

86

each strand:

EcoRI is produced by certain strains of thebacterium Escherichia coli as a defenseagainst infection by the DNA of bacterio-phages. The enzyme severs the viral DNAand thus prevents its propagation in thebacterium. (To avoid self-destruction, thebacterium also produces another enzymethat modifies its own DNA so that it cannotbe severed by EcoRI.) other bacterial speciesproduce restriction enzymes that recognizeother sequences. For example, HindIII, therestriction enzyme used by our collaboratorsat Livermore for their phase I libraries, isproduced by the bacterium Hemophilus in-fluenzae. Like EcoRI, HindIII recognizes asix-base sequence (–A–A–G-T–T--T–)and makes staggered cuts.

Complete digestion of the chromosomalDNA with EcoRI (or HindIII) produces a setof relatively small flagments. We can esti-mate the average length of these fragmentsby assuming a random distribution along thechromosomal DNA of equal numbers of thefour bases A, T, G, and C. Then the prob-ability of any six successive bases being theEcoRI (or HindIII) recognition sequence is(1/4) 6, which implies an average of about4100 base pairs between cleavage sites, (Anintact DNA molecule in the largest humanchromosome (number 1) is about 2.5 X 108

base pairs in length. )We, and our collaborators at Livermore,

use DNA from the bacteriopbage Charon

2 I A as a cloning vector. that is. as a mediumcapable of accommodating foreign DNA seg-ments for replication. Charon 21A is a modi-

DNA molecule is about 42,000 base pairslong, terminates with short stretches ofsingle-stranded DNA, and includes a singleEcoRI cleavage site.) The DNA moleculesextracted from the phages are cut by EcoRIinto two "arms.” The cut ends of these armsare then stripped of- a phosphate group (for areason to be explained later) by treatmentwith the enzyme alkaline phosphatase.

The next step is fairly simple, at least intheory. We mix the two DNA preparationsand let nature take its course—nature in thiscase meaning the tendency for single strandsof DNA to “stick together” (become linkedby hydrogen bonds) at regions with com-plementary bases, Since the cutting action ofEcoRI has equipped both the chromosomalDNA fragments and the vector arms withjust such sticky ends. aggregates consisting of’two arms linked by a (single) chromosomalDNA fragment will form. (Herein lies thereason for gentle extraction of the DNA fromthe chromosomes: physical breaks do notresult in sticky ends. ) These aggregates arcthen converted into intact DNA moleculesby treatment with the enzyme DNA ligase,which establishes a covalent bond at thebreak along each strand. The result is a set ofrecombinant DNA molecules—that is, mol-ecules of phage DNA carrying human DNAfragments. (To increase the probability offorming recombinant aggregates. thechromosomal fragments are mixed with anexcess of arms. Therefore, many of the ag-gregates formed will consist simply of twoarms. But DNA ligase cannot repair the


Genes by Mail

Extract DNA

Digest with EcoRI

Mix and Treat

with DNA Ligase

Assemble Phages

I

Infect Bacteria

Collect Library of Cloned Phages


III. CLONING THE: DNA

DNA in such aggregates because of the miss-ing phosphate groups, So our foresight in theprevious step reduces the background ofnonrecombinant DNA molecules in the finallibrary.)

Next we assemble Charon 21A phages,drawing on proteins produced by mutantvarieties of the phage for the necessary struc-tural components and on our supply of re-combinant DNA molecules for the geneticcomponent. This step unavoidably in-troduces a deficiency in the resulting library:those recombinant DNA molecules contain-ing fragments longer than about 9100 basepairs are too large to be encased in the headsof the phages, and therefore the library willlack that portion (about 33 mass percent) ofthe chromosomal DNA residing on suchfragments. It is for this reason that we andour collaborators at Livermore use differentrestriction enzymes and thus produce li-braries based on different sets of chromo-somal DNA fragments. The expectation isthat the DNA missing from one laboratory’sversion of a library will be found in the other,

The final step in making the libraries, thatof cloning the recombinant DNA molecules,is accomplished by seeding the phages onto a“lawn,” or monolayer, of an appropriatestrain of E. coli, When a phage injects itsrecombinant DNA molecule into such a cell,the biological apparatus of the cell issubverted to production of new phages. Thecell ruptures when between one and twohundred new phages have been produced,and these then infect surrounding cells. Theprocess continues. resulting in formation of aplaque, or clear area, in the translucent lawnof bacteria. In each plaque arc some one toten million infective phages, each containingan exact copy of the recombinant DNAmolecule carried by the original infectingphage.

These vastly multiplied recombinantDNA molecules, then, constitute a library-conveniently and safely packaged within theproteinaceous armor of phages. This con-struction of a library puts us at the startingline. Its application to problems of clinicaland fundamental significance lies ahead. ■

what genes or combinations of genes areinvolved; in fact, we haven’t come close tosettling even the nature-versus-nurture issue.And even when a single defective gene isknown to be the culprit—as it is in certainkinds of mental retardation—developing thetechniques for introducing the normal geneand getting it expressed properly will take along time.HILDEBRAYD: Yes, routine clinical ap-plication of gene therapy is a long way off,especially in light of the ethical and legalquestions it raises, But in the process ofironing out those questions, society shouldfocus on the extraordinary benefits to bederived from the technology rather than its“Brave New World” implications.MOYZIS: I think other developments in

molecular genetics pose more immediateconcerns. For instance. the availability of-

probes specific to the X and Y chromosomeshas made it very easy to determine the sex ofa fetus. Since even in "sexually liberatcd”western countries the preference for having amale child first is still strong. routine de-termination of fetal sex could conceivablylead to dramatic changes in the ratio of thesexes. The ethical. social, and economic im-plications of this possibility need immediateconsideration.SCIENCE: Are issues such as these heavy onyour mind as you go about your research?CRAM: I don’t think so. The goals of basicresearch arc information and understanding.It’s up to society to determine how those areapplied. Take the photoelectric effect, forexample. Its applications are mainly viewedas beneficial, although a few may be regardedby some as deplorable. But a scientist doesn’thide his head in the sand in the face ofinformation that can possibly be applied in aharmful way.MOYZIS: I would second that. You can’t beparalyzed because a bit of knowledge can beused inappropriately. And 1 repeat that mostof the worrisome aspects of gene therapy area long way down the road. A major problemthat must he solved before it can be appliedto many diseases is how to introduce thegood genes into only the appropriate CellS of

87

a patient. If the defect is in a gene that isexpressed, say. only in liver cells. extra copiesof this gene in other cells might cause a greatdeal of mischief.

I am somewhat amazed, however, at thespeed with which the medical community ismoving toward gene therapy. I’m aware,though, that clinicians confronting patientshave a different viewpoint than we do. It’s noparticular virtue on the part of people inbasic research to be more cautious.

I think the patient and the patient’s familyshould have a large say in deciding aboutexperimental treatments. If a child of minehad Lesch-Nyhan syndrome, I’d be willing totake a chance on gene therapy. Terminalcancer patients are taking similar chancesright now with other experimental therapies.HILDEBRAND: An indirect form of genetherapy has already been tested clinically.Patients with a form of thalassemia weretreated with a drug that would—it washoped—reactivate the fetal hemoglobingenes. The drug worked, but since it is apotential carcinogen, the ultimate result maybe initiation of tumors or other defectsthrough inappropriate activation of othernormally silent genes. Of course, all thera-peutic regimes have the risks of unwantedside effects. Gene therapy presents the specialrisk of possibly altering the human gene pool.But so do chemotherapy and radiation ther-apy.

Direct gene therapy seems imminent. andpublic interest is high, even though the ge-netic diseases now being considered as can-didates for this method of treatment are ex-tremely rare. The Office of Technology As-sessment has published a background paperon gene therapy, and the NIH’s Recombi-nant DNA Advisory Committee has ap-pointed a Working Group on Human GeneTherapy. The legal problems may take a longtime to clear up, though.DEAVEN: Yes, but that can change rapidly.About six years ago Margery Shaw, Directorof the Institute of Medical Genetics at theUniversity of Texas in Houston, gave a collo-quium here at Los Alamos on the legalaspects of genetics. At that time the fetus had

absolutely no legal status in this country or inother western countries. Last year she gavean update of that talk at the annual meetingof the American Society of Human Genetics.She reported that in parallel with scientificdevelopments in recognizing forms of fetalabuse, such as overconsumption of alcoholor other drugs by the mother, principles arebeing established concerning the legal statusof the fetus and its right to be born in a stateof good health. My feeling is that a similarthing will happen with gene therapy, and thatby the time gene therapy is a reality we willhave arrived at legal guidelines for its ap-plication.SCIENCE: Do you think society turns toscientists-for ethical guidance as well as tech-nical expertise?CRAM: Yes, and I think most of us acceptthat role, But the scientific community canlose credibility because even scientists caninterpret the same technical data differently,This happened when policies for radiationexposure were being established.MOYZIS: That’s because we’re human be-ings just like everybody else. Although wemay be more aware of the possible applica-tions of our research than the average person,most of us are not particularly qualified tospout off about the ethics of those applica-tions. Whether or not I favor abortion orgenetic engineering of humans is in someways based on my scientific expertise but inother ways is not. We’re often called upon atcocktail parties to discuss these sticky issues,but I tend to feel that most scientists shouldbe asked for nothing more than scientificinput. Most scientists aren’t any more objec-tive than anybody else when it comes tojudgments that involve emotions and prej-udices.DEAVEN: Yes, but I do think scientistsshould spend some time thinking about theimplications of their research and shouldactively participate in the formulation ofethical and legal guidelines for society. Somehave altered their careers in this direction,I’ll mention Margery Shaw again. After ac-quiring an M.D. and having a successfulcareer in human genetics, she went back to

school and got a degree in law because shesaw the legal implications of new develop-ments in human genetics. We need morepeople like that.HILDEBRAND: This is all very interesting,but perhaps we should retreat to a moredirect application of the libraries—that ofdeveloping more sensitive methods for de-tecting mutations.MOYZIS: Yes, we badly need a means ofdetecting damage to human DNA caused bylow-level exposure to mutagens, It’s not hardto detect the damage caused by high-levelexposure. Ionizing radiation, for example,causes hundreds to thousands of double- andsingle-strand breaks in the DNA molecules

of a single cell at exposures over 1000 rads.But very few people ever get that kind ofexposure. Most of the concern, medical andlegal, is about the effects of low dose rates. Isexposure to fallout from the test of a nuclearweapon in Nevada some thirty years ago, orto Agent Orange in Vietnam, the cause ofsome of today’s cancers? Does living in LosAlamos at an altitude of 7500 feet cause ahigher incidence of spontaneous abortions?Who knows? It’s very difficult to make theseassessments.

Currently, the incidence of mutations isdetermined from the observable changeswithin a small number of genes, most com-monly a single gene. Since a single gene isOnly about one-millionth of a DNA mol-ecule, the assays are not very sensitive. Butwith a battery of hundreds or perhapsthousands of gene probes pulled from thelibraries, a much larger fraction of the DNAcan be assayed for damage. A wealth ofcytogenetic data indicates that some regionsof certain chromosomes are more susceptibleto damage. Obviously, probes for those re-gions would be of particular value.DEAVEN: Los Alamos is already workingon detecting mutations induced by low-levelradiation by assaying what are known astandem repeats for RFLPs.MOYZIS: Yes, and although our results onirradiated cells are very preliminary. they’revery encouraging. We’re working with therepetitive sequences in human DNA. which


Genes by Mail

make up about one-quarter of the total.Many of these repetitive sequences arethought to be involved in defining chromo-some structure or in switching genes on andoff. They often jump to new locations in theDNA, altering the gene sequence they moveinto. We've cloned and characterized over 75percent of the repetitive DNA sequencespresent in human DNA. Most of these se-quences occur about a hundred thousandtimes per genome. and about a third of themoccur in tandem, one right after another.With a single RFLP probe for such a tandemrepeat, we can sample many orders of magni-tude more DNA for mutations than withprobes for a few genes. With this increase insensitivity Susan Atchley, a graduate stu-dent in the Laboratory’s Genetics Group, hasbeen able to detect damage to DNA at amuch lower level of radiation exposure thanhas been reported previously. But it appearsthat most of the observed changes in theDNA disappear with time following irradia-tion. Presumably the changes arc being re-paired. These preliminary results have led usto facetiously throw around the idea—although maybe it’s not so facetious-that a little bit of’ radiation may be a goodthing in the sense that it keeps the repairenzymes induced to such a level that they canmore readily cope with larger doses.

The repair we are seeing has relevance tothe long-standing question of whether or notsuch a thing as a threshold for damage exists.Perhaps there is a dose rate below which theDNA repair enzymes can handle the damage,even with continuous exposure.CRAM: If repair is blocked. what levels ofdamage can be detected?MOYZIS: We haven’t done that experimentyet.DEAVEN: With the current techniques forassaying genetic damage, large groups of ex-posed people must be studied to get goodstatistics, Very few such groups arc available.In fact. despite a massive effort since 1945,the data emerging for the largest group—theHiroshima and Nagasaki populations-arequestionable in terms of genetic significance.The power of the newer techniques is that

they can be applied to small populationsbecause they provide so much more informa-tion per person.

MOYZIS: The recent incident in Mexico, inwhich some two hundred people were ex-posed to cobalt-60 pellets from a junkedradiation therapy machine, gave us a newpopulation for study. but those people gothorrendous doses.DEAVEN: One problem with most of theavailable study populations is that the ex-posure is not known with certainty. We can

estimate the exposure by determining levelsof chromosome damage, but this method isnot very accurate, especially if the chromo-some analysis is not done immediately. Datafrom laboratory experiments, such as thoseBob is talking about, should provide bettercorrelations between damage and dose.CRAM: Is there so much variation in restric-tion fragment lengths among individualsthat detecting mutations requires a controlsample for each individual?MOYZIS: We’ve found that the variation

The success of the Los Alamos contribution to the Gene Library Project is due tothe competence and enthusiasm of a large number of people. Pictured here areKevin Albright, Marty Bartholdi, Bill Bentley, Nancy Brown, Evelyn Campbell,Doug Chritton, Scott Cram, Larry Deaven, Ed Hildebrand, Paul Jackson, TedLobb, Mary Luedemann, John Martin, Linda Meinke, Julie Meyne, and BobMoyzis. Jim Jett, Jon Longmire, and Chris Munk were unavailable for thephotograph.


depends on the sequence. For damage esti-mates we’ve concentrated on sequences thatseem not to vary in the population at large.Preliminary variation was estimated byisolating DNA from blood samples of anyindividual who was unfortunate enough towalk past my lab. The Life Sciences Divisionhere has enough people with different ethnicbackgrounds to make those estimates mean-ingful,

Repetitive sequences that do vary in thepopulation are also interesting. Ones thatmap to particular chromosome locations willmake exquisite probes for those chromo-somes and will be useful in prenatal diag-nosis, since orders of magnitude less fetalDNA will be needed for analysis. Ed and Iare currently searching for such sequences inthe libraries.

HILDEBRAND: Another application ofprobes from the libraries is investigatingpermanent changes in the genome on a largertime scale, for example, in evolution andspeciation,

DEAVEN: Yes. Using a battery of probesthat represent an ordered arrangement ofhuman genes, one can look for patterns ofconstancy and of change in the genomes ofplants and animals. That sort of informationshould give some insight as to just what thelimits are on rearrangement of the genome.For example, if a group of genes has re-mained linked closely together over hun-dreds of millions of years of evolution, that’sa pretty good sign that any perturbation ofthat conserved linkage is deleterious or evenlethal, Taxonomists might be able to usesuch information as the basis for a betterdefinition of speciation.

MOYZIS: It might also help solve thedilemma of cross-species extrapolation. Therelevance of evidence from an animal studyto humans can always be questioned. But if itis known what genes or organizations ofDNA are unique to, say, mice and to humansand common to both, estimates of risk as-sessment from animal studies will be onfirmer ground.

DEAVEN: Pedigree analysis will also benefit

from the libraries. Pedigree analysis is thetracing of a genetic trait through generationsof a family to determine whether the gene forthe trait is dominant or recessive andwhether it is located on one of the sexchromosomes or on an automsome. For thisjob geneticists have been using phenotypicevidence of the presence of a gene. But somegenes don’t provide unambiguous pheno-typic evidence. The use of RFLP probes forsuch genes will make pedigree analysis easierand more accurate.

MOYZIS: In the near future enough probesfor RFLPs will be available that it will bepossible to assign paternity with 100 percentcertainty.

CRAM: I believe a test like that was done onthe pandas at the Washington zoo to find outwhether the female was impregnated by hermate or by artificial insemination with spermfrom a panda in England.

MOYZIS: I didn’t realize that, but similartechniques are used routinely to distinguishdifferent strains of plants. DNA might cometo replace fingerprints for certain identifica-tion purposes.

SCIENCE: Has production of the librariesbeen fairly routine or full of surprises?

VAN DILLA: Dealing with life processesand living things practically guaranteessurprises. An example is our experience withhybrid cells. The first hybrid cell line weworked with was just a dream. The cells wereeasy to grow and. as advertised, containedthree human chromosomes, which were easyto sort. So it came as a rude surprise thatother hybrid cell lines didn’t perform nearlyso well, One was supposed to contain humanchromosome 1, but after the cells weregrown, we couldn’t find any to sort. We hadknown that hybrid cells have a habit of losinghuman chromosomes as they multiply buthad been lulled into a false sense of securityby the behavior of the first cell line.

Even so routine a process as chopping upDNA with a restriction enzyme gave us asurprise. We were getting a cutting efficiencyof only about 1 percent even though we wereusing a standard restriction enzyme whose

cutting efficiency was well documented to beclose to 100 percent. At first we blamed a badbatch of enzyme, but that idea didn’t checkout. We still don’t know exactly what theproblem is, but it’s been solved by an addi-tional purification of the DNA extractedfrom the chromosomes. So we assume theenzyme was being poisoned by some con-taminant the chromosomes pick up duringsorting,

The logistics of producing the librariesgave us a surprise too. When we switched tosorting with the high-speed flow cytometer inthe spring of ’84, the poor cell farmers weresuddenly faced with keeping up with its tentimes greater appetite for chromosomes.That shouldn’t have come as a surprise, butit did, probably because we were too busywith other aspects of the project to think hardabout the manpower needed to keep thehigh-speed sorter happy,SCIENCE: Have you enjoyed participatingin the Gene Library Project?HILDEBRAND: Definitely yes, but therehas been an element of frustration because itis taking a much longer time than we antici-pated. We’re still adapting to the reality thatthis project is a long-term research and devel-opment effort with a defined endpoint. TheLaboratory has vast experience in dealingwith long-term experiments in physics. Lifescientists, however, have traditionally beenaccustomed to experiments that can be com-pleted in a period of days, weeks, or months,To a physicist what we call an experimentwould be just one step in a much largerprocess.MOYZIS: Just a coffee break. But that doesbring up a good point, For the Life SciencesDivision the Gene Library Project is a majorundertaking in terms of the time necessary toget to the starting point of actually using thelibraries. The DOE may have been right inthinking that Livermore and Los Alamoswere the only places where enough biologistswould talk to each other long enough to getthe project going. Physicists are more ac-customed to projects involving lots of peo-ple. Biology is still very much done by oneguy with his test tube at his bench.


Genes by Mail

DEAVEN: Group research in biology is go-ing to become more and more common,though. As more and more sophisticatedtechnologies come into use, the breadth ofexpertise needed in a research effort getsgreater. This project is certainly an exampleof that trend. Between twelve and fifteenpeople, representing a variety of specialties,are involved on a day-to-day or week-to-week basis.CRAM: The point that Ed made about thedifference between biologists and physicistsis worth emphasizing. We can do an experi-

ment sometimes in an afternoon, butphysicists often spend years on an experi-ment. A biologist might publish six papers ayear, but physicists doing an experiment atan accelerator might publish a single paperwith ten authors every couple of years. Ithink the Gene Library Project is an examplethat crosses that barrier.MOYZIS: Biology is still such a wide openfield with so much left to learn that we canget significant, publishable results in a rel-atively short time. A number of people in-volved in this big project are going to have

Acknowledgments

The following people have made substantial contributions to the organization and administration ofthe National Laboratory Gene Library Project: Mark W. Bitensky, Leader of the Life SciencesDivision at Los Alamos National Laboratory; Brian D. Crawford, now Director of the MolecularBiology Program at Long Island University; M. Duane Enger, now Head of the Department ofZoology at Iowa State University; Mortimer J. Mendelssohn, Associate Director of the Biomedicaland Environmental Research Division at Lawrence Livermore National Laboratory; David A.Smith, Office of Health and Environmental Research, U. S. Department of Energy; Carleton C.Stewart, Leader of the Pathology Group at Los Alamos National Laboratory; Robert Wagner,Consultant at Los Alamos National Laboratory; and Ronald A. Walters, a member of the staff of theAssociate Director for Chemistry, Earth and Life Sciences at Los Alamos National Laboratory.

Further Reading

James D. Watson, John Tooze, and David T. Kurtz. Recombinant DNA: A Short Course. ScientificAmerican Books, 1983; distributed by W. H. Freeman and Company.

Alan E. H. Emery. An Introduction to Recombinant DNA. Chichester, England: John Wiley & Sons,1984.

Maya Pines. “In the Shadow of Huntington’ s.” Science 84, May 1984, pp. 32-39.

Office of Technology Assessment. Human Gene Therapy—A Background Paper. Washington,District of Columbia: U.S. Congress, Office of Technology Assessment, OTA-BP-BA-32, December1984.

little holes in their curricula vitae becausethey felt it was so important. Not many of thebiologists at universities would have beenwilling to devote two years or so to gettingthese libraries made.DEAVEN: Working on something so dif-ferent as this project has certainly been an in-teresting challenge. Sometimes though, andthis is probably true of all of us, I dream ofbeing on the other side of the fence—opening my mail, unwapping a little vial, andactually doing some research with alibrary. ■

Yvonne Baskin. “Doctoring the Genes.” Science 1984, December 1984, pp. 52-60.


Documents

Genes - Federation of American Scientists · PDF fileGenes Participants discuss the Gene research on the genetic endowment) ect and species.,.. its Mlial~. .,,,’ ., ... By convention